Management of a High Arctic River

(1)

NTNU Norwegian University of Science and Technology Faculty of Engineering Department of Geoscience and Petroleum

Martin Andreas Solbakken LøvaasManagement of a High Arctic River

Martin Andreas Solbakken Løvaas

Erosion and Sediment Transport in Longyearelva, Svalbard

Master’s thesis in Arctic Geology Supervisor: Bjørn Frengstad

Co-supervisor: Aga Nowak and Lena Rubensdotter May 2021

Photo: Arne T. Pedersen (Svalbard Museum)

Master ’s thesis

(2)

(3)

Martin Andreas Solbakken Løvaas

Management of a High Arctic River

Erosion and Sediment Transport in Longyearelva, Svalbard

Master’s thesis in Arctic Geology Supervisor: Bjørn Frengstad

Co-supervisor: Aga Nowak and Lena Rubensdotter May 2021

Norwegian University of Science and Technology Faculty of Engineering

Department of Geoscience and Petroleum

(4)

(5)

I

Abstract

The High Arctic town, Longyearbyen (78⁰13’N, 15⁰38’E), is rearranging its infrastructure as a consequence of increasing geohazards and awareness. After establishing permanent scour and flooding mitigation (sills and riprap) in Longyearelva, several institutions found it interesting to enhance the understanding of the local glaciofluvial system draining the 22.2km² catchment. Therefore, the objectives for this thesis were to; - Assess the scour and flooding mitigation over the first ablation season since completion. - Quantify discharge, erosion, sediment transport and investigate the associated sources. - Contribute to the long-term monitoring in Longyearelva.

Fieldwork was conducted from June 5^th to September 15^th, covering most of the 2020 ablation season in Longyearelva. Water-stage and discharge rating curves were established from a water pressure transducer and point measurements of discharge from slug injection of diluted salt. Water samples were acquired using an automatic ISCO-pump, and suspended sediment concentration (SSC) was determined gravimetrically and used to calculate suspended sediment yield (SSY). Bedload transport was monitored using coloured passive tracers (pebble to large cobble). Geomorphological mapping of the moraines and the fluvial system was based on remote sensing data with relatively high spatial- and temporal resolution combined with field observations.

The hourly hydrograph illustrates an average discharge of 1.5m³/s and peaks up to 8.6m³/s. The peaks correlate with the record high air temperatures in late July. The highest recorded SSC at 24.1g/l coincides with the rising limb of the hydrograph on July 25^th and indicates flushing of easily available sediments. The specific SSY was 1866t/km²/yr, and both SSC and SSY were considerably higher than in comparative catchments - erosion and active layer detachments in the moraine could have increased the sediment input combined with construction work. Bedload transport capacity is at least large cobbles, as transport of the largest passive tracers was documented. The sedimentation dam with an initial volume of 30 000m³ had to be excavated to increase the capacity after the late July Flooding. The riprap remained intact with only minor damages downstream Sill 8. Several sills collapsed or subsided, some developed scour holes, and a few remained intact.

Lithology, construction work, and channelized water arguably affect the ground thermal regime and thus the mechanical strength of the ground around the sills – which could explain the subsidence and collapse. The scour mitigation limited a general channel degradation despite some scouring holes and transport of passive tracers over Sill 3 and 18.

Baseline data for further studies were acquired as planned and contributed to the established monitoring program (RIS ID 11641). It is believed that the system needs time to achieve an equilibrium between capacity and available sediments before the full effect of the scour mitigation can be revealed. An analysis of thermal data in the fluvial channel and the moraines are considered highly interesting for further research.

(6)

II

(7)

III

Sammendrag

Som en konsekvens av økende fokus på geofarer gjennomføres en storstilt omstilling av infrastruktur og boligmasser i Longyearbyen, Svalbard (78⁰13’N, 15⁰38’E). Permanent erosjon- og flomsikring av Longyearelva ble ferdigstilt høsten 2019, etter en fireårig byggeperiode. Flere organer viste interesse for å igangsette et forskningsprosjekt som kunne resultere i en bedre forståelse av de rådende forholdene i det 22,2km² store nedbørsfeltet rundt Longyearbyen. Denne masteroppgaven har dermed følgende problemstillinger; - Evaluere det nylig ferdigstilte erosjon og flomvernet. - Bidra til langsiktige observasjoner i Longyearelva, der vannføring, erosjon og sedimenttransport skal kvantifiseres og de respektive kildene skal identifiseres.

Feltarbeidet ble gjennomført i perioden fra 5. juni til 15. september og dekket det meste av smeltesesongen i 2020. Vannstand og vannføring ble målt med henholdsvis en vanntrykksensor og punktmålinger med injeksjon av oppløst salt, før forholdet ble uttrykt i likninger. Vannprøver for kvantifisering av transport av suspendert materiale ble innhentet med hjelp av en automatisk ISCO-pumpe mens bunntransport ble overvåket med bruk av fargede markører (kornstørrelsene grus – blokk). Geomorfologisk kartlegging av morenene og det fluviale systemet ble basert på fjernanalyse og feltobservasjoner, med god oppløsning både romlig og tidsmessig.

Kalkulering av kontinuering vannføring ga et gjennomsnitt på 1,5m³/s i 2020, der toppene i hydrografen på 8,6m³/s korrelerer med de rekordhøye temperaturene i den siste uken av juni. Høyeste målte konsentrasjon av suspendert materiale (KSM) var 24,1g/l (25.juni) og det spesifikke resultatet for transport av suspendert materiale (TSM) var 1866t/km²/yr.

Ekstremverdiene for KSM samsvarer med de første oppsvingene i hydrografen, noe som indikerer en utvasking av lett tilgjengelig sediment. Både KSM og TSM er betydelig høyere enn i relativt tilsvarende nedbørsfelt - betydelig erosjon i morene kan ha økt tilførselen av materiale i tillegg til bygningsarbeid i kanalen. Kapasiteten for bunntransport i Longyearelva er små blokker eller større, ettersom selv de største markørene ble transportert 80m. Sedimentasjonsdammen som hadde kapasitet til 30 000m³ ble fylt og måtte tømmes i etterkant av flommen sent i juli. Plastringen av elvevollene viste bare tegn til skade nedstrøms Bunnbånd 8, som ellers viser tegn til kollaps og store setningsskader, noe som også ble dokumentert på andre bunnbånd. Litologi, byggeprosessen og kanalisering av vannet er elementer som antas å øke temperaturen i permafrosten og dermed redusere den mekaniske styrken i grunnen rundt bunnbåndene - en mulig forklaring på de skadene som har oppstått. Bunnbåndene har til en viss grad oppnådd målet om å hindre bunnsenkning, til tross for utbredte erosjonsgroper og transport av passive markører over bunnbånd 3 og 18.

Resultatet av målingene fra 2020 bidrar som planlagt til datagrunnlaget for videre arbeid i et langtidsperspektiv (RIS ID 11641). Den fulle effekten av flomvernet ligger noen sesonger frem i tid, ettersom det antas at systemet trenger tid til å oppnå likevekt mellom transportkapasitet og tilgjengelig materiale. Analyse av termisk data fra elvekanalen og morene er anbefalt for videre forskning.

(8)

IV

(9)

V Preface

After five years at NTNU, the journey has come to an end and is expressed in the following 100 pages or so. This 60 ECTS Master Thesis within Arctic Geology at NTNU and UNIS is the outcome of 16 weeks of fieldwork in the Longyearbyen area during the summer and autumn of 2020. It is now more or less one year since the first run along the Longyearelva at the beginning of my field season, and the final deadline approaches. I was fortunate to be one of the few Guest Master Students at UNIS during the challenges caused by Covid- 19, and I am genuinely grateful for the opportunity.

I would like to thank those who have contributed during the process. My supervisor, Bjørn Frengstad (NTNU), for guidance and discussion. My co-supervisor Aga Nowak (UNIS) and Lena Rubensdotter (UNIS/NGU) for involving me in the project, encouragement during fieldwork and valuable input during the process of writing these pages. Anders Bjordal from NVE, for help during fieldwork and discussion of the scour and flooding mitigation and Longyearbyen Lokalstyre (w/Kjersti Olsen Ingerø) for letting me tag along on inspections.

The Logistical Department at UNIS provided equipment and expertise for effective fieldwork, an essential factor for the successful outcome of the fieldwork. Spending three months in Longyearbyen to gather data was possible thanks to the financial support from both the Research Council of Norway (Arctic Field Grant) and UNIS (funds for Guest Master Students, Arctic Geology).

Thanks to friends and family in Trondheim, Svalbard, and at home for offering mental diversions. Thanks to SvalbardButikken for selling the world’s most fantastic chocolate, my go-to companion for fieldwork and coffee breaks; Marabou Apelsin Krokant. SteinKubben and Kaffidrekkarlauget made my time as a student in Trondheim memorable. Although, the coffee breaks lasted longer than they should, and the amount of chocolate consumed would frighten any diet expert out there.

Martin Solbakken Løvaas Trondheim, 15.05.2021

Cover Photo

The cover photo is provided by Svalbard Museum (historical photograph library) and Arne T. Pedersen (photo). Bulldozers at work in Longyearelva around 1950-60, making a deeper thalweg to control the water and sediment transport away from infrastructure. The location is approximately where Veg 501 is today.

(10)

VI

(11)

VII

1 Introduction

1.1 Motivation and objectives

Erosion and sediment transport in a fluvial system is closely linked to the discharge and the force of the water flowing through a channel (Fergus et al., 2010). The establishment of permanent scour- and flooding mitigation is recently completed in Longyearbyen, and the small town is rearranging its infrastructure in the light of increased geohazard awareness. Long term data-series of erosion and discharge is sparse on Svalbard (Nowak et al., 2021; Sund, 2008), and it is a crucial need for an up-to-date and representative dataset for the Longyeaelva river. A collaboration between Longyearbyen Lokalstyre (LL), the Norwegian Water- and Energy Directorate (NVE), and the University Centre in Svalbard (UNIS) were therefore initiated. This thesis is an outcome of the collaboration and will during the 2020 ablation season focus on the following objectives:

- Quantify the discharge and sediment yield from the Longyearelva catchment.

- Investigate the sources for discharge and sediment input.

- Assess the adequacy and limitations of the recently established scour- and flooding mitigation.

- Contribute to the long-term monitoring in the catchment and identify additional topics to address over the coming years.

The tasks were tackled with comprehensive fieldwork from early June to mid-September.

Fieldwork involved establishing a hydrological monitoring station, measurements of sediment transport, and geomorphological mapping of the moraines and the constructed channel. See Chapter 3 Methodology for further description.

The next subsections introduce the historical and present social situation in Longyearbyen, followed by a description of the Svalbard climate, Longyearelva catchment and a summary of the continuous attempts at managing the local river.

1.2 Historical background and present situation

Longyearbyen is located in the middle of Spitsbergen (78⁰13’N, 15⁰38’E), the largest island in the Svalbard archipelago (NPI, 2020b) see Figure 1. Longyearbyen is the main settlement on Svalbard with 2400 inhabitants (SSB, 2020), including infrastructure expected in modern society (SSB, 2016). Longyearbyen was founded in the early 20^th century as a small coal mining community (Arlov, 1994), and the remnants of the first infrastructure are still a part of the local scenery. For instance, a couple of the old cableway ramps can be recognized along Longyearelva (Figure 7 and Figure 8). Coal mining was the primary industry for several decades. However, the demand for coal began to decrease in the second half on the century (Arlov, 2020) and tourism is now the number one source of income for the community (Elliassen, 2020; SSB, 2016). As of 2021, the only active coal

(18)

2

mines are Gruve 7 and one mine in Barentsburg, a Russian mining settlement. Gruve 7, operated by Store Norske Spitsbergen Kulkompani (SNSK), provides coal for the local power plant (SNSK, 2020).

The social development from mine workers living in barracks to a fully developed family society led to an increasing need for residential areas in the 1980s (SSB, 2016; Arlov, 2020). A new residential area called Lia developed to meet the rising number of families, in contrast to the barracks in Nybyen and Sverdrupbyen (Arlov, 2020). Elvesletta was also of interest for building purposes in the 1990s (Sværd, 1996).

SNSK had responsibility for everything in Longyearbyen during the middle of the 20^th century, but responsibilities for infrastructure such as roads, pipelines and buildings were gradually transferred to the Norwegian Government and Longyearbyen Community Council (Longyearbyen Lokalstyre, LL). LL was aligned as a local democracy in 2002 (SSB, 2016).

The transition from a coal mining society to the present Longyearbyen was enhanced when The University Center in Svalbard (UNIS) was established in 1993 (UNIS, 2020). Research and education are now one of the cornerstones in the economy (SSB, 2016).

The infrastructural planning and housing situation took a dramatic turn during Christmas in 2015. A snowstorm on the 19^th of December triggered a snow avalanche that crushed eleven houses in Lia, caused several injuries and the loss of two lives (DSB, 2016).

SvalbardPosten (the local newspaper) posted an article from the one-year memorial that emphasizes the social scar made by the avalanche in such a small community (Røsvik, 2016). In February 2017, another avalanche was triggered in the same area, hitting two houses, but it caused only material damages (Landrø et al., 2017). The remaining houses in Lia were decided to be demolished in the following years. The need for safe housing facilities became precarious, and the Elvesletta area was again in focus as a potential site for establishing residential areas (LL, 2019). The river needed, therefore, to be permanently controlled to prevent erosion and flooding.

(19)

3

Figure 1: Overview map of Longyearbyen and the Svalbard archipelago. Detailed Longyeardalen area with relevant labeling based on location mentioned in the text. Annotated from TopoSvalbard, NPI (2020).

(20)

4 1.3 Geological setting

1.3.1 Bedrock lithologies

Tectonic plate movement caused a collision of Svalbard (Eurasian) and Greenland, a foreland basin developed as the tectonic plates diverged after the collision of Svalbard (Eurasian) and (Müller and Spielhagen, 1990). The foreland basin is now known as The Central Spitsbergen Tertiary Basin (CTB) and accumulated sediments in the Paleogene period (Dallmann, 2015). The Longyeardalen stratigraphy consists of horizontal, slightly southward dipping, sedimentary rocks from the CTB (Major and Nagy, 1972), see Figure 2. The overall trend in the sedimentary lithostratigraphy is a general regression, with several trans- and regressions during Paleocene and Eocene. The bottom units in the area marine sand- and siltstone with beds of shales from the Cretaceous, whereas the top units are mostly terrestrial sandstone (Dallmann, 2015; Müller and Spielhagen, 1990).

The Carolinefjellet Formation (Cretaceous) holds alternating marine sand and mudstones but is only exposed in some road cuts and mostly covered with sediments from the younger stratigraphy (Dallmann, 2015). The Firkanten Formation, with sandstones, conglomerates, and a coal seam, marks the transition (hiatus) from the underlying Carolinefjellet Formation to the Paleocene epoch (Major and Nagy, 1972; Müller and Spielhagen, 1990).

Steel et al. (1981) further describe the Firkanten Formation, divided into Todalen- and Endalen Member as followed. The foremost holds the rich and easily accessible coal layers, the very reason for the Longyearbyen settlement. The latter makes out the first pronounced cliff-forming sandstone bodies in the valley sides, see Figure 2 and Figure 3.

Grumantbyen and Hollendardalen Formation are described by Dallmann (2015) as strongly bioturbated sandstone, visible as the uppermost cliff-forming unit and form the flat plateaus, Sverdruphammeren and Gruvefjellet, see Figure 2. Steel et al. (1981) differentiates the two Formations and describes the Hollendardalen Formation as sandstone, wedge-shaped, which disappears towards the northeast in the CTB with beds of shale. It is, therefore, likely that it is the Grumantbyen Formation that is exposed Longyeardalen, even though the geological map from the Norwegian Polar Institute (NPI, 2020a) displays them as one. Steel et al. (1981) further describes the Grumantbyen Formation as strongly bioturbated greenish sandstone.

Helland-Hansen (1990) described the alternating silt and sandstones in the Battfjellet Formation, based on outcrops in the cliffs south in the Longyeardalen. The underlying dark shales belong to the Frysjaodden Formation. On top of the Battfjellet Formation lays the thick, terrestrial sand and siltstone-rich Aspelintoppen Formation (Helland-Hansen, 1990).

The formation can be found in the uppermost parts of the area, such as in the vertical cliffs in the south-western corner of Longyearbreen, see Figure 2 (Helland-Hansen, 1990; Müller and Spielhagen, 1990). Etzelmüller et al. (2000) characterize the clastic sedimentary rocks in the area as mechanically soft, easily eroded, and fine-grained, factors that are crucial for grain size distribution and further sediment transport.

(21)

5

Figure 2: Geological map annotated from NPI (2020a), illustrating the geological units around Longyearbyen.

The lithology is in general, mechanically weak sedimentary rocks. The ‘’pickaxe and shovel’’ indicate the old coal mines and the Todalen Member.

(22)

6 1.3.2 Quaternary geology and geomorphology

Glaciers are a typical feature of Svalbard due to the high latitude and Arctic climate. The glacier inventory is versatile, with large ice caps and smaller valley glaciers (Dallmann, 2015). Hagen et al. (1993) reveal that glaciers covered more than 60% of the land area.

However, relatively recent research by Nuth et al. (2013) indicates a 57% glacier coverage and claims that a general negative glacier mass balance caused the decrease. With the NCCS (2019) climate report in mind, it is expected that the glaciated area today is even lower than the findings in Nuth et al. (2013). The glacier inventory in the surroundings of Longyearbyen consists of an abundance of relative small valley- and cirque glaciers (Hagen et al., 1993).

Humlum et al. (2003) define the Longyearbyen area as a zone with continuous permafrost (see chapter 2.2.1 Permafrost and ground thermal regime). Glaciers in a zone of continuous permafrost will, according to Ødegård et al. (1992), cause the small glaciers to be cold-based, meaning that they are frozen to the ground, relatively immobile and less abrasive. The fact that Humlum et al. (2005) documented in situ vegetation at the base of Longyearbreen illustrates that at least it is immobile, which also concurs with the findings in Etzelmüller et al. (2000). The permafrost thickness around Longyearbyen has not been studied in detail. Still, Humlum et al. (2003) describe thicknesses around 100m along the coast and 4-500m further inland for the area in general. The permafrost in Longyeardalen is described by Gilbert et al. (2019) as saline with high ice content.

The landscape around Longyearbyen (see Figure 3) consists of wide U-shaped valleys, with braided rivers draining the glaciers. Plateau mountains flank the valleys, influenced and controlled by the horizontal stratigraphy, and the steep mountainsides are covered with landslide deposits (Lied and Hestnes, 1986). The harder quartz-rich sandstone withstands the physical erosion better than the softer shales and therefore stand out as pronounced cliffs (Lied and Hestnes, 1986), see Figure 3. The area has minimal vegetation and an absence of plants with considerable rooting systems (Lied and Hestnes, 1986).

The sediment thickness in the Longyeardalen varies, although the different boreholes provide only point data, and the complete picture is uncertain (Instanes and Rongved, 2017). Investigations by Gilbert et al. (2018) in the northern part of Longyeardalen show marine clays and a gradually coarsening upward into deltaic deposits. The marine limit is around 70m above the current sea level due to isostatic lifting (Gilbert et al., 2018;

Instanes and Rongved, 2017). The fact that a marine shell found at 3.8m depth in a borehole in Lia by Berggren and Finseth (2019) supports the conclusion of isostatic lifting and change in sea level. The high salinity can consequently originate from the influence of seawater and marine sediments.

(23)

7

An increasing sediment thickness is to be expected towards the centre of the valley, although local variations may be caused by jagged bedrock morphology (Instanes and Rongved, 2017). Gregersen (1995) conducted a series of drillings along Longyearelva (from Veg 501 to the outlet, see Figure 1) and concluded that spatial variation of sediment thickness could be the result of shifts in the erosional- and depositional environment over time, a plausible explanation which also was discussed by Lied and Hestnes (1986). The uppermost 2-3m alongside the Gregersen (1995) drilling profile consists of coarse gravel.

Underneath is a 3-4m thick, wedge-shaped layer of sand, which thins out and disappears halfway downstream the profile. Below this sand layer was a laterally persisting layer of clay, 3-8 m thick, inter-bedded with layers of silt. In the outer edge of the drilling profile, another layer of sand was documented in the deepest part of the sediment sequence.

Figure 3: Annotated drone pictures to illustrate the landscape around Longyearbyen. A) An overview looking northwards, downstream the Longyearelva. B) looking westward at Longyearbreen and the moraine. C) Looking southwards at Larsbreen and the moraine.

(24)

8

Gregersen (1995) also detected high salinity with an increasing concentration towards the outlet in Adventfjorden. Lied and Hestnes (1986) conducted a series of sieving tests to investigate the geotechnical properties of the sediments. However, the material in the fluvial channel was only briefly described, as it was believed to be unproblematic for engineering properties due to the low content of fines (clay, silt, and fine sand). A single sieving test resulted in 95% sand, gravel, and cobble while the remaining 5% were fine sediments.

Nårstad et al. (2018) described a 24m deep borehole east of Elvesletta as a coarser top layer of gravel, followed by alternating sand and clay and clay with layers of silty sand towards the bottom, without reaching bedrock. Pedersen (2018) described sediment thickness of 20-28m down to bedrock at another borehole at Elvesletta, with similar stratigraphic content as Nårstad et al. (2018). However, even courser sediments in the top layer were described, possibly due to the closer proximity, and thus influence, of the glaciofluvial system.

Geotechnical investigations down to 15m around the UNIS building (see Figure 1) did not reach the bedrock. Investigations by Gregersen and Tuft (1994) and Gilbert et al. (2019) documented a consolidated top layer of coarse gravel, followed by more silty sand and clay with some coarser sections towards the bottom of the drillings. In contrast, LNS Spitsbergen reached bedrock at 11-13m in Sjøområdet, after drilling through nothing but clay (LNSS, 2016). Drilling at the more elevated area at Skjeringa (Figure 1) shows relatively coarse material (sand and gravel) with only 8m of sediment above the bedrock (Instanes and Rongved, 2017).

Sediment thickness and origin change from the central valley towards the steep mountainsides, with more input from mass movement events and less fluvial sedimentation (Lied and Hestnes, 1986). Eckerstorfer et al. (2013) describe the slope the morphology and sediments as snow avalanche colluvial fans, with sediment material reflecting weathering of the local lithostratigraphy, a description similar to the findings in Lied and Hestnes (1986). The more low-angled alluvial fan at the mouth of the small side- valley Vannledingsdalen reaches far out on the valley floor and is related to repeated slush- avalanche activity at Haugen, see Figure 1.

The periglacial areas in front of the Longyearbreen and Larsbreen glaciers are described by Etzelmüller et al. (2000) as ice-cored moraines with a 0.5-1.5m surface layer of mixed sediment, see Figure 3. The eastern margin of the Larsbreen moraine complex has later been categorized as avalanche-derived rock glaciers, see Figure 3 (Humlum et al., 2007).

The till in the Larsbreen moraine is considerably more fine-grained than the till at the Longyearbreen moraine. Etzelmüller et al. (2000) argue that this is due to the higher topographic position of Larsbreen, which results in sediment input from mechanically weaker lithological units, such as the shales from the Frysjaodden Formation (see Figure 2). In contrast, the lower situated Longyearbreen has eroded into coarser lithologies and more resistant lithologies, such as Firkanten- and Battfjellet Formation. Longyearbreen and thus the moraine receive additionally input through rockfall and snow avalanches from the Aspelintoppen Formation (see Figure 2). The moraines described above are constantly eroded by the glacier meltwater streams from the glaciers and are thus directly connected with the Longyeardalen glaciofluvial system. A simplified summary of the sediment distribution in Longyeardalen is illustrated in Figure 4.

(25)

9

Figure 4: Schematic illustration of the general deposits and sediment thicknesses in Longyeardalen based on a basic interpretation of the data presented in the text.

(26)

10 1.4 Hydrological conditions

1.4.1 Svalbard weather and climate

The Norwegian Centre for Climate Services (NCCS, 2019) report clarifies that climate changes are pronounced in the High Arctic. Svalbard and the surrounding ocean show the most extensive temperature increase and loss of sea because of the climate changes (Isaksen et al., 2017; NCCS, 2019). The rise in air temperature is most noticeable during the winter months (Isaksen et al., 2017; Wawrzyniak and Osuch, 2020).

Isfjorden, a massive fjord in the middle of the Spitsbergen island (see Figure 1), allows ocean currents to bring warm ocean water inland from the west coast, resulting in a warmer climate than expected so close to the North Pole (Walczowski and Piechura, 2011).

Eckerstorfer and Christiansen (2011) suggest a ’High Arctic maritime Snow Climate’ to best describe the climate in the vicinity to Longyearbyen, and a long-term dataset from Hornsund documents how the ocean currents contribute to relative humid conditions along the west-coast (Wawrzyniak and Osuch, 2020). Precipitation on Svalbard is historically sparse with large spatial variations (Førland and Hanssen-Bauer, 2003; Isaksen et al., 2017) and Wawrzyniak and Osuch (2020) demonstrate how late autumn precipitation events are common. However, a warmer climate alters the precipitation patterns, e.g., increasing volumes and rain during the winter is more frequent (NCCS, 2019). Strong winds cause snowdrift, and the measured precipitation might not reflect the actual snow coverage in the catchment, as large cornices and uneven spatial distribution of the snow have been documented by Hancock et al. (2018)

The nearest weather station is located at Svalbard Airport (28 m.a.s.l) 4km northwest of the Longyearbyen city centre, see Figure 1. The monitoring station has been functional since 1976, operated by the Norwegian Metrological Institute, MET Norway, and precipitation and temperature data is displayed in Figure 5. The mean annual air temperature over the last 30 years (1991-2020) is -4.7⁰C, and the average yearly precipitation over the same period is 202mm (MET, 2021).

Figure 5: Historical weather data from the meteorological station at Longyearbyen Airport, data from Norwegian Metrological Institute (MET, 2021).

MET (2021)

(27)

11 1.4.2 Longyearelva catchment description

The Longyearelva catchment consists of glaciers, mountains, and infrastructure, unlike most of the catchments on Svalbard, see Figure 6. The total catchment area is 22.2km², including two glaciers covering 5.7km². Etzelmüller et al. (2000) mapped Longyearbreen and Larsbreen to be 2.7km²and 3.0km²,respectively, and described them as cold-based, although a small patch of temperate ice was discovered in the uppermost western corner of Longyearbreen. However, considered the climate situation described in NCCS (2019) and the effect on small valley glaciers, the thickness and area of the glaciers are likely to have decreased. Longyearbreen meltwater stream receives a limited contribution from Platåbreen, which drains partly through Tverrdalen. The meltwater streams from Larsbreen and Longyearbreen glaciers confluence near Nybyen and forms Longyearelva, 3.3km from the outlet in Adventfjorden.

The gradual expansion of Longyearbyen has restricted the river into an artificial channel in the middle of the valley. The old aerial photographs by the Norwegian Polar Institute (NPI, 1936) (Figure 7A) show the natural state of Longyearelva as it filled the whole valley floor back in 1936. Reconstruction of the historical photograph illustrates the current situation in 2020, demonstrated in Figure 7B.

Figure 6: Longyearelva catchment covering 22.2km², ranging from above 1000 m.a.s.l to the outlet in Adventfjorden, including two glaciers and Longyearbyen. The watershed is calculated using a 5x5m DTM and adequate hydrology modelling-tools.

(28)

12

Figure 7: Illustrating changes in Longyeardalen over the last 90 years. A) Aerial oblique photo from NPI (1936), showing the natural state of Longyearelva and the early settlement. B) Drone photo from 2020, replicating the photo in A. Note the cableway ramp (1) and the cableway station (2) marked with red circles for reference points.

(29)

13 1.4.3 Previous research in Longyearelva catchment

Discharge and sediment transport in the catchment has previously been the topic in several research projects; Etzelmüller et al. (2000) investigated glacier characteristics and sediment transport in the catchment over two seasons, 1993-94, and Yde et al. (2008) studied the hydrochemistry in the meltwater from Longyearbreen. Furthermore, two master projects, Grønsten (1998) and Riger-Kusk (2006), have covered discharge and fluvial sediment transport over one or two seasons. The neighbouring catchment towards the east, Endalen catchment, drains into Isdammen (the water reservoir for Longyearbyen). NVE has monitored discharge and sediment transport in Endalselva over several years during the 1990s (Bogen and Bønsnes, 2003; Sund, 2008).

1.5 Measurements and previous initiatives

The white paper ‘’Svalbard’’ from the Solberg Government in 2016 states that extensive funding for permanent geohazard mitigation measures in Longyearbyen is a part of the national budget (St.meld 32 (2015-2016), 2016). The white paper further states that NVE has the responsibility to evaluate geohazards on behalf of the Department of Justice and Public security, and LL will be responsible for maintenance. Geohazard mitigations must now meet the strict regulations on the Norwegian mainland. A more long-term perspective regarding geohazard mitigation was therefore needed.

1.5.1 Development of geohazard mitigation

Geohazard mitigation has been an ongoing process since the establishment of the town.

Although previous initiatives seem to be short-termed, and more permanent measures tend to follow dramatic events. A deadly slush avalanche hit the residential area at Haugen in 1953 (Larsen, 2016) (see Figure 1). Despite countermeasures such as reinforced embankment and extensive usage of bulldozers to clear the narrow Vannledningsdalen, the area was struck once again in 1989 (Larsen, 2016). Snow cornices breaking and triggering snow avalanches a real danger in Longyeardalen (Vogel et al., 2012). SNSK used explosives to remove the cornices when coal miners lived in barracks in Nybyen (Larsen, 2016). However, evacuation of the residents is now preferred, as the snow avalanches are still a hazard (Bårdseth, 2021).

1.5.2 Mitigation measurements in Longyearelva

The gradual expansion of infrastructure on the valley floor required extensive usage of bulldozers to control the river and protect infrastructure, such as the cableway-ramps, see Figure 8 (Pedersen and Svalbard Museum, 1960; Hoseth and Daae, 1996; Bjordal and Hoseth, 2017). SNSK increased the usage of bulldozers following a large flood in 1964 and planned a sedimentation dam next to Nybyen after consultations from NVE over the following years; however, the plans were never commenced (Hoseth and Daae, 1996).

SNSK and NVE considered more permanent solutions for flooding mitigation again in 1989, but the usage of bulldozers continued (Øvereng, 1989; Hoseth and Daae, 1996). Elvesletta area was investigated for the development of residential and commercial buildings as early as the 1980 and ‘90s (Lied and Hestnes, 1986; Gregersen, 1995; Hoseth and Daae, 1996).

Preliminary flooding calculations and plans for controlling the river were initiated in the mid-90s (Hoseth and Daae, 1996; Sværd, 1996). However, it would take another 20 years before the plans were put into action.

Extensive usage of bulldozers continued until NVE initiated the construction of more permanent erosion and flooding measures in 2016. The action plan by Bjordal and Hoseth (2017) states that the levees needed scour-protection once the water was channelized,

(30)

14

and a sedimentation dam should be built between Huset and the school (Figure 1) to control the sediment supply (see Figure 9). Shutting off the sediment supply could lead to channel degradation, and sills were therefore built. Construction work was completed in 2019, and the assessment of the mitigation is hence an essential part of this thesis. The different measurements and specific constructions are further explained in Chapter 2.6 Hydrological engineering and scour protection.

Critical infrastructure, such as bridges and culverts, has also been reinforced. Old culverts have been replaced with bigger weirs and spillways, such as around Veg106, Veg501, and Veg600 (Larsen, 2016; Bjordal and Hoseth, 2017). The bridge at Veg503 is missing, however, extra scour protection is already in place (Bjordal and Hoseth, 2017). Larsen (2016) illustrates how the new bridges have been built with reinforced fundaments, scour protection and are designed to withstand a 200year-flood event while limiting the influence on the ground thermal regime by allowing natural heat flux between the ground and the air.

Figure 8: Crucial use of bulldozers to protect infrastructure along Longyearelva river dates to the 1950’s (Pedersen and Svalbard Museum 1960). In this case is a bulldozer is used to keep the river from eroding the fundaments of the cableway ramps used for transportding coal from the mines to the harbour.

(31)

15

Figure 9: Timeline for construction work (building sills and riprap) in the river by NVE since 2016 and placements of all the sills and the sedimentation dam. Riprap is continuous on both sides from the sedimentation dam to Veg 600.

(32)

16

(33)

17

2 Theoretical background

2.1 Arctic conditions

The arctic conditions influence the geological processes and features in the morphology.

Catchments with glaciers are normal on Svalbard (Hagen et al., 1993), but Longyearelva, strongly influenced by human activity, is one of a kind on the archipelago. Thus, a combination of literature and technics from a global and local perspective is needed to cover all relevant theoretical aspects.

2.2 Heat flow and thermodynamics

Some basic terms must be defined to understand the heat flow through the sediments in Longyeardalen and the coupling between water, air, and thawing of the uppermost section of sediments.

2.2.1 Permafrost and ground thermal regime

Svalbard and, therefore Longyearelva, are in a zone of continuous permafrost (Humlum et al., 2003), which has important implications for both hydrology and engineering.

Permafrost is defined based on the ground temperature, which cannot exceed 0⁰C for two consecutive years (Andersland et al., 2003). The perennially frozen ground holds distinct mechanical properties compared to the thawed counterpart, as the mechanical strength increases with a frozen soil-skeleton and permeability will be limited or absent (Andersland et al., 2003).

An important aspect is the yearly fluctuating air temperature and the response in the ground thermal regime, illustrated with the trumpet curve, see Figure 10. The uppermost section of the ground where temperatures rise above 0⁰C for a period of the year is defined as the active layer, illustrated in Figure 10. The air temperature affects the ground temperature even deeper than the active layer, known as the depth of zero annual amplitude. Instanes and Rongved (2017) show active layer thickness at 1.5m in Longyearbyen city centre and depth of zero annual amplitude at 5-10m in several boreholes around the area, whereas Bjordal and Hoseth (2017) documented active layer thickness in Longyearelva at 2.5m. The active layer reaches a maximum depth late in the ablation season (September) (Andersland et al., 2003).

The permafrost depth is controlled by the geothermal gradient, historical climate conditions, thermal properties in the ground. Temperature readings from a borehole close to the Longyearbyen airport document permafrost thickness at 22-39m (NGU, 2007), in contrast to the general 100m thickness in coastal areas (Humlum et al., 2003). The active layer thickness is controlled by the heat exchange between the ground and the air.

Increased insulation from, e.g., snow or infrastructure can hence slow down the refreezing, while infrastructure or vegetation can provide shadow and reduce the warming during the summer. The temperature in the active layer can further change the depth of zero annual amplitude and thus affect the properties of the permafrost (Andersland et al., 2003).

Buildings in Longyearbyen are built on stakes to allow natural air circulation and heat exchange or with artificial cooling in the foundation, both to limit the effect on the ground thermal regime (Hestnes et al., 2016; Larsen, 2016). Increased permafrost temperature and a thicker active layer decrease the strength of the foundations and can cause settling damage (Andersland et al., 2003). Reestablishing permafrost after construction work and thus achieve the intended foundation has been a problem for the Svalbard Global Seed Vault (Statsbygg, 2019).

(34)

18 2.2.2 Thermal conductivity

Thermal conduction is the transfer of kinetic energy from an area of high energy to an area of low energy (Andersland et al., 2003). Thermal conductivity is further described as how efficiently a material transmits energy - the ability to transfer heat increases with the dry density and saturation of soils. Materials have various conductivity, e.g., the mineralogy of soils or bedrock will affect the thermal characteristics. Water has a high conductivity compared to air and, likewise, with metamorphic rocks compared to sedimentary rocks (Labus and Labus, 2018). The relative conductivity (C) between material represented in Longyeardalen: Cmineral > Cice > Cwater > Cair (Woo and Xia, 1996). Subsequently, the heat transfer efficiency decreases over the ablation season as the ground ice starts to melt, water drains, and the pores are filled with air. Andersland et al. (2003) define the ratio of thermal conductivity and density of soil as the thermal diffusivity, in other words, how fast heat is transferred through a material. Frozen soil will have a higher diffusivity compared with a thawed sample of the same soil. The above will influence the heat transfer through the soil as the uppermost layer start to thaw, and the active layer thickens (Andersland et al., 2003).

Figure 10: Trumpet curve annotated from Andersland and Ladanyi (2003) to fit the conditions in Longyearelva catchment. Illustrating the thermal regime in the ground with the annual variation based on surface

temperature. Cite specific considerations are based on Instanes and Rongved (2017) and Humlum et al (2003).

(35)

19 2.2.3 Heat capacity

Heat capacity is the amount of energy (heat) needed to raise the temperature in a given sample by 1⁰. Different components in soils, e.g., water, mineralogy, and air, will have different heat capacities. Adding the specific values from each material determines the net heat capacity of a given soil or deposits (Andersland et al., 2003).

2.2.4 Heat flow in soils

The general heat flow in sediments depends on the thermal conductivity and the temperature differences. The energy exchange between the atmosphere and the ground can be expressed through the energy balance. The penetration of surface energy will gradually decrease with increasing depth until a point where the temperature in the ground is stable all year round (dept of zero annual amplitude, see Figure 10).

2.3 Hydrology in the Arctic

Discharge in Arctic rivers is typically limited to the ablation season, May to September (Killingtveit, 2004), and is generally frozen during the remaining months of the hydrological year (October 1^st to 30^th of September). However, as a result of a warmer and wetter climate, the timing of initial discharge and freeze-up is changing. (NCCS, 2019; Hestnes et al., 2016).

Snowmelt occurs early in the ablation period, typically May-June, leading to the first peak in discharge. The accumulation of snow in the catchment is therefore important as water is temporarily stored during winter. Less snow means less insulation for the permafrost and glaciers, hence will the glacial melt start earlier in the summer if the snow cover already is gone. For the rest of the ablation season (June-September), discharge and glacial coverage in the watershed are closely coupled through air temperature and glacial melt (van Pelt et al., 2019). Minor floods late in the ablation season tend to correspond with precipitation (Killingtveit, 2004; Nowak and Hodson, 2013). However, snowmelt- induced floods are not the most prominent hazard, according to Hoseth and Daae (1996), but it is instead the high temperatures and precipitation during July-September that cause the most severe floods and related engineering difficulties.

Water percolating through the active layer can pick up solutes from the sediments and increases electrical conductivity (EC). Groundwater has, therefore, a higher EC compared to meltwater from clean snow and clean ice (Yde et al., 2008). The EC in Arctic rivers increases late in the ablation season as the meltwater contribution declines and the active layer thickens (Yde et al., 2008) – which illustrates the coupling of ground thermal regime and the hydrological conditions in a catchment.

2.3.1 Water balance in Longyearelva

As described above, several factors influence the discharge in a catchment, and the water balance can be expressed with Equation 1 (Killingtveit, 2004).

Equation 1 𝑃𝐴− 𝑄𝑆− 𝑄𝐺− 𝐸𝐴± ∆𝑀 = 𝜀

Where PA is the input through precipitation measured in mm, QS refers to the surface discharge and QG to the groundwater flow, both measured in cubic meters per second

(36)

20

(m³/s). EA is the actual evaporation, and ∆M is water storage. Ꜫ is the error term and should ideally be zero based on the accuracy of the other parameters (Killingtveit, 2004).

An important factor regarding precipitation described in Killingtveit (2004) and Nowak and Hodson (2013) is the elevation distribution in a catchment (which can be illustrated by a hypsographic curve). Precipitation can change from rain to snow with increasing elevation due to decreasing temperatures. Strong winds can cause under-catchment in gauging stations, and snow depth in a catchment may not correlate with the recoded precipitation due to redistribution (Dingman, 2015). Observations from, e.g., Wawrzyniak and Osuch (2020) document increased precipitation, and climate models (e.g., Hansen et al., 2014;

Bintanja and Andry, 2017; NCCS, 2019) show a continuing increase in precipitation and a warmer and wetter Arctic in the future, especially during winter months.

QS, hereafter called discharge, is measured or computed, e.g. based on the water stage- discharge relationship (Dingman, 2015). Groundwater flow or seepage can be more challenging to measure. Groundwater flow in permafrost areas is relatively poorly understood (Neilson et al., 2018), however as stated by Nowak et al. (2021), the topic has received increased attention. The investigations of pingos in the Adventdalen valley illustrate sub-permafrost groundwater flow in the area (Hodson et al., 2020), and the understanding of permafrost groundwater is improving. Groundwater flow will increase due to a warmer climate and permafrost degradation across the Arctic, according to models from Bense et al. (2009). The fact that geotechnical investigation by Pedersen (2017) documented water seepage at a depth of 4-7m at Elvesletta indicates some groundwater in the catchment, although the exact contribution is yet to be investigated in detail.

As discussed in Killingtveit (2004) and Dingman (2015), the potential evaporation may be significantly higher than the actual evaporation given the midnight sun. Despite this, the actual evaporation on Svalbard is considered to be minimal due to the geology, sparse vegetation, low precipitation, and cold temperatures (Killingtveit, 2004). Evaporation can potentially increase and thus also precipitation due to the observed increased temperatures and less sea ice (Bintanja and Andry, 2017).

Water storage (∆M) is a vital part of the water balance in glaciated catchments on Svalbard as the air temperature and glacier ablation control discharge (van Pelt et al., 2019). Both snow accumulation and ground ice contribute to the total water storage in a catchment.

However, van Pelt et al. (2019) show that the contribution from snowmelt and groundwater is limited compared to the glacier melt. Accumulation of snow and ice could keep the glacier mass balance in equilibrium, and change in storage could be neglected. However, the glaciers on Svalbard are shrinking in time with a warmer climate, and ∆M is hence positive (NCCS, 2019; Nowak et al., 2021).

2.3.2 Glacial hydrology

The high proportion of glacial coverage on Svalbard constitutes most of the water storage across the archipelago and is thus highly relevant for the water balance in glaciated catchments (van Pelt et al., 2019). The thermal regimes of glaciers are of high importance for seasonal trends, glacial runoff, and sediment yield (Hodson et al., 1997; Hodson and Ferguson, 1999). Previous literature concludes that the combination of small, thin glaciers and continuous permafrost makes a typical Svalbard polythermal or cold-based as size decreases (Ødegård et al., 1992; Björnsson et al., 1996; Hodson et al., 1997). The glaciers on Svalbard are generally shirking (NCCS, 2019), and a shift from polythermal to cold- based is documented, e.g., at Austre Brøggerbreen (Nowak and Hodson, 2014). In cold- based glaciers, like the ones in Longyearelva catchments, supra-glacial drainage is more

(37)

21

important than subglacial drainage (Etzelmüller et al., 2000). The linkage between the thermal regime in glaciers and sediment transport has been widely investigated (e.g., Hodson et al., 1997; Etzelmüller et al., 2000; Bogen and Bønsnes, 2003; Hodgkins et al., 2003), and temperate or polythermal glaciers tend to produce both higher discharge and sediment transport rates compared to the cold-based counterparts.

2.4 Erosion

Charlton (2007) defines fluvial erosion or scouring as the relationship of transport capacity and sediment supply in a stream. Erosion can be understood as a loss of material and channel degradation. Erosive forces and flow regimes are further described.

2.4.1 Gravel- and cobble bed rivers

The term ‘’gravel-bed river’’ is a collective term used in literature for describing rivers where gravel is the median grain size, e.g., in Kociuba (2014) and Laronne and Carson (1976). A combination of the grain size classes defined in Wentworth (1922) and the typical Norwegian grain size chart presented in Fergus et al. (2010) is shown in Figure 11. Bunte and Abt (2001) presents principles for the classification of rivers with different dominating grain sizes. Classification of sediments is typically based on a sieving test and plotted in a logarithmic grain size distribution curve. The mechanical strength of the local lithology is crucial as it will reflect the grain size distribution, based on how easy sediments are weathered and crushed during transport. However, the term gravel-bed river seems to be widely used in the literature regarding high Arctic rivers. Sediment sizes referred to in this thesis are based on the classification presented below.

2.4.2 Fluvial morphology

The valley floors on Svalbard are typically covered by glaciofluvial outwash plains (Dallmann, 2015) or sandar (sandur in singular), where the shape and size of the valley influence the morphology (Boothroyd and Ashley, 1975; Rudberg, 1988).

Braided rivers are recognized by the numerous bars and channels, which are repeatedly flooded and migrating laterally. The bars tend to be longitudinal and with imbrication, special within larger clasts (Nichols, 2009). Braided channels in a sandur are common due to high bedload transport, channel aggregation, and ever-changing discharge (Krigström, 1962; Boothroyd and Ashley, 1975). Krigström (1962) further describes the change in morphology as the distance from the glacier snout increases. Forking, or braiding, becomes more frequent further downstream. Boothroyd and Ashley (1975) highlight the effect of Figure 11: Grain size chart for classification of mineral sediments based on Wentworth (1922) and Fergus et al. (2010). The Norwegian chart disregards granule and pebble and categorizes gravel from fine to coarse.

The transition from cobble to boulder is at 200mm in the Norwegian system, not 256mm as it is in the Wentworth (1992) chart.

(38)

22

the surrounding topography as length, width, and gradient of the available area will control the fluvial morphology.

Anastomosing rivers are distinguished by interfingering channels with bars or flood plains in between (Nichols, 2009). Both channels and bars are more stable than in braided river systems, with less channel aggradation and lateral migration.

Arctic rivers remain frozen for most of the year, given the cold climate and can thus be defined as ephemeral rivers. Ephemeral rivers are characterized by how discharge is limited to events, such as rain and snow or glacier melt (Nichols, 2009; Dingman, 2015).

2.4.3 Flow regimes

Specific flow regimes in a stream control the erosive forces and thus effects the mitigation measurements needed. Firstly, flow regimes are divided between laminar or turbulent flow.

The former is defined based on the parallel movement of water molecules and no mixing within the water column (Charlton, 2007; Nichols, 2009). The latter is defined based on water particles move in all three dimensions with a net downstream movement, resulting in a highly efficient mixing within the water column (Charlton, 2007; Brooks et al., 2012).

Grain size, relief, geology, and human activity affect the channel characteristics and hence the flow regime in the river. A laminar flow interacting with an obstacle in the channel will result in a turbulent flow, or eddies, downstream of the obstacle. Turbulent flow causes more uplift and thus higher stress on the bed material (Charlton, 2007; Brooks et al., 2012) see Figure 12. Kay (2008) differentiates between subcritical- and supercritical flows based on the Froude number. A Froude number greater than 1 refers to a supercritical flow and subcritical when the number is below 1. A supercritical flow has high energy and can be very erosive (Charlton, 2007; Kay, 2008). A breaking wave develops directly in transition between sub- and supercritical flow, known as a hydraulic jump. Given a supercritical flow, the speed of the water will be greater than the wave speed.

2.4.4 Erosive forces

Water moves due to gravity and shifts from potential energy to kinetic energy. Kinetic energy allows the water to perform work and apply shear stress on the wetting perimeter

Figure 12: Illustration of the forces working on the riverbed, annotated from Fergus et al., (2010). The flow is laminar until the roughness of the bed causes disturbance and thus a turbulent flow. The flow can turn supercritical as the water velocity increases as the white flow-arrows are closing in over the grains.

(39)

23

(Charlton, 2007). The wetting perimeter is known as the zone where water interacts with the channel. Charlton (2007) further presents the Du Boys Equation, see Equation 2, and defines the shear stress [N/m] in a fluvial channel.

Equation 2 𝜏₀= 𝜌𝑔ℎ𝑆

Where τ0 is the average shear stress applied on the channel, (ρ) is the density of the water, (g) acceleration due to gravity, (h) refers to water depth and (S) is the gradient of the channel. The shear stress will increase with an increasing inclination (S) and water depth (h). A steeper channel will thus produce higher shear stress and potentially more erosion than its counterpart. Increasing the wetting perimeter and thereby decrease the depth (h), will reduce the shear stress.

The Shield parameter is utilized to calculate the threshold shear stress for initiating movement on a particle with a given dimension (Charlton, 2007). The Shield parameter is expressed with Equation 3.

Equation 3

𝜃

_𝑐𝑟

=

^𝜏^𝑐𝑟

𝑔(𝜌𝑠−𝜌)𝐷

Where θcr is the critical bed shear stress or Shield parameter, τcr is the shear stress.

Acceleration due to gravity (g), density of the sediment (ρs), density of water (ρ), and D refers to the grain size dimensions. A small particle with low density will be more exposed to erosion and transport than the opposite particles, based on the Shield parameter. The shape and orientation of each grain in relation to the flow direction are critical, in addition to size and density (Self et al., 1989).

2.5 Sediment transport

Flowing water is the prominent transport medium for sediment through the Longyearelva catchment (Lied and Hestnes, 1986). Sediment transport in a fluvial system is divided into two main patterns of movement: in suspension or as bedload, based on the interaction of water, sediment, and channel bed (Nichols, 2009). Hjulström (1935) illustrated the relationship between water velocity and grain size for erosion, transport, and deposition with the Hjulstrøm Curve, see Figure 13.

Sediment transport will mirror sediment sources in the fluvial system. Bank erosion can cause undercutting, potentially collapse, and consequently increased sediment input. The mechanical strength of local lithology is important for the capability to resist erosion (Brooks et al., 2012). Factors such as permafrost and glacial characteristics are also contributing under Arctic conditions. Glaciers that abrade the bedrock can induce considerably higher erosive forces than flowing water. The active layer thickness is of high interest concerning sediment supply, as frozen sediments are more stable and harder to erode (Andersland et al., 2003; Instanes and Rongved, 2017).

(40)

24 2.5.1 Suspended sediment transport

Based on the Hjulstrøm curve (see Figure 13), fine particles (<63µm) will be transported even on low flow velocities and are typically kept in suspension in the water column (Hjulström, 1935). Suspended transport is caused by turbulent forces exceeding the gravity working on the grain (Brooks et al., 2012). In highly turbulent and powerful water, even fine sand can be transported in suspension (Kay, 2008). Threshold values for entrainment of fine-grained sediments can be relatively high due to cohesive forces (Brooks et al., 2012). Particularly when compared to the forces needed for transporting the grains once suspended. Suspended sediment can be transported over long distances, depending on the flow regime.

2.5.2 Bedload sediment transport

Courser sediments, e.g., sand, pebbles, and cobbles, are transported as bedload as the uplift from turbulent waters does not exceed the gravitational force on the grains. Bedload is characterized by rolling and saltation along the channel floor (Nichols, 2009). Larger grains will require more energy for initiation motion based on the Shield parameter and the Hjulström Curve. A pebble in motion can collide with other particles and thus increase the stress and further increase the bedload. Pitlick et al. (2008) discuss the effect of larger particles protecting the underlying fines in a channel, creating an armouring layer. Coarse sediment (>sand) can occur in imbrication patterns, increasing the anchoring effect leading to higher critical shear stress for entrainment.

Figure 13: The Hjulstrøm Curve, logarithmic curve for illustrating the relation of water speed and erosion- transport-deposition regime in the fluvial channel with respect to the grain size (Hjulström, 1935).

(41)

25

2.6 Hydrological engineering and scour protection

Constructions and engineering in conjunction with waterways aim to protect infrastructure from potential hazards. Mitigation measures will minimize the effect of an event, whereas preventional measurements reduce the chance for a hazard to occur. The NVE has developed detailed guidelines for waterway constructions on mainland Norway, described in detail by Fergus et al. (2010). Building material is typically blasted or quarry rock, but due to the poor quality of the local lithology, granitic gneiss had to be shipped from mainland Norway (Bjordal and Hoseth, 2017). Alternative methods to blasted rock is to use gabions (Hoseth and Daae, 1996) or concrete blocks (Reid and Church, 2015). A gabion is a cube or a mat of wire net, holding rocks in place and can be used in the same ways as building bricks. A gabion does not require the same rock strength and size as riprap, and Sværd (1996) mentioned the opportunity to utilize local resources in gabions. Still, it was not recommended due to challenging maintenance.

2.6.1 Sedimentation dam

A sedimentation dam or pool aims to obstruct further bedload in the channel (Fergus et al., 2010). By constructing a dam and thereby decrease the energy of the water, bedload will start to accumulate. The capacity of the dam will inevitably decline, and maintenance (excavation) depends on the accumulation rate (Fergus et al., 2010). Maintenance is mentioned in the building plan by Bjordal and Hoseth (2017), although a detailed schedule is missing due to uncertain accumulation rates.

Based on the building plans from NVE (Bjordal and Hoseth, 2017), the sedimentation dam was built as follows (see Figure 14 for illustration). The barrier is made up of a 2m high weir across the channel, with large boulders (100-150cm) facing upstream. Smaller boulders (50-100cm) were used for the apron downstream the crest. The apron aims to reduce the energy of the water flowing over the crest and thus avoid scouring holes. The large boulders allow water to percolate through while sufficiently decreased the energy to allowing bedload to settle. The levees directly upstream were reinforced and riprapped to avoid undercutting and collapse. The dam has an initial capacity of thirty thousand cubic meters of sediments.

Figure 14: Annotated picture of the Sedimentation dam, riprap on both flanks to avoid undercutting. The dam consists of large boulders of blasted granitic gneiss, with an apron of smaller boulders downstream. The large boulders allow water to percolate through, while reducing the energy and allow bedload to settle.

Management of a High Arctic River - Erosion and Sediment Transport in Longyearelva, Svalbard

Martin Andreas Solbakken Løvaas

Management of a High Arctic River

Erosion and Sediment Transport in Longyearelva, Svalbard

Master ’s thesis

Martin Andreas Solbakken Løvaas

Management of a High Arctic River

Erosion and Sediment Transport in Longyearelva, Svalbard

Master’s thesis in Arctic Geology Supervisor: Bjørn Frengstad

Co-supervisor: Aga Nowak and Lena Rubensdotter May 2021

Norwegian University of Science and Technology Faculty of Engineering

Department of Geoscience and Petroleum

Abstract

Sammendrag

Contents

1 Introduction

2 Theoretical background

𝜃

=